wega virtual breakfast: Artificial intelligence
in GxP environments

Date:

March 23, 2021 08:30 - 10:00

Link to the recording:

Video recording of the talks

Abstract

While emerging technologies such as artificial intelligence (AI) hold great promise for improving life science processes in many ways, existing GxP regulations and validation guidelines remain largely unchanged.
This event will discuss the current situation of the regulatory framework for AI, present concrete use cases for the application of AI in the regulated GxP industry including the validation approach used, and discuss the challenges and opportunities related to the introduction of AI-based systems.

Talk 1: AI validation in the pharmaceutical industry

AI is different. A trained machine learning algorithm has an input and an output, everything in between is black box by nature.
This presentation shows how trust in such a black box can be achieved by certain additional validation steps. The current status of the work of an ISPE GAMP DACH working group will be described. The aim of the group is to produce a guideline that brings pharmaceutical companies, suppliers, consultants and authorities to the same level of understanding when it comes to AI in GxP.
The presentation will be completed by a real use case. In a chromatography process, a biofluid is separated into API and other by products. Today, an operator decides when to start and stop the extraction of the API. He looks at the time series curves of the process and decides with his knowledge when to start the extraction.
The presentation shows how this manual step could be replaced by an algorithm trained with historical data of the process and how this model could increase the yield of the process. It will be shown how the steps to build trust in AI from the GAMP group are implemented in this project.

 

Talk 2: Validating Intelligent Automation Systems in Pharmacovigilance

  • Introduction
  • Presentation of  available materials of "Intelligent Automation Opportunities in Pharmacovigilance" from Transcelerate page.
  • References and other interesting materials - quick walkthrough (in terms of challenges and opportunities):
    • Celgene, Pfizer
    • FDA x3
    • Explainable AI
    • Pre-trained models (deep learning)
  • Final stories:
    • How to validate an online AI
    • Building chatbots for patients, lessons
    • Trust in AI

The Speakers:

Rolf_Blumenthal

Rolf Blumenthal

Rolf Blumenthal is a senior consultant with many years of MES experience in the pharmaceutical industry. He retired in 2017 and now works as an independent MES consultant (continues to support Werum). In his last position Mr. Blumenthal led the product development of Werum Manufacturing Execution Systems (MES) Software Suite PAS-X for 10 years. He has outstanding expertise in the management of large and international MES software projects for the pharmaceutical and biotech industries (e.g. Bayer, Boehringer, Roche, Novartis, Merck, J&J, Amgen, MSD, Takeda, Astra Zeneca, Astellas). In 2017, he founded a new Werum division to gain the additional value of data from the pharmaceutical industry, and he is still working for this new division. It is the first time he can combine his pharmaceutical knowledge with his skills in statistics, algorithms and numbers.

Rolf Blumenthal is a recognized expert with extensive knowledge and practical experience in cGMP, FDA requirements, GAMP guidelines and computer validation.

Mr. Blumenthal is one of the initiators of the GAMP D-A-CH Forum and a member of the Steering Committee. In this role, he managed Special Interest Groups for a review of whether agile programming fits GAMP V and presented several times at the annual GAMP V conferences.

Pawel_Dobracki

Pawel Dobracki

Pawel Dobracki is an IT Expert Software Validation at Roche Global IT SC.
He joined Roche more than 10 Years ago bringing his vast expertise, great teamwork spirit, knowledge-sharing attitude and tons of new ideas. His previous experience started in the Testing domain.
He loves brainstorming, thinking about the future and optimization. One of his recent involvements includes co-authorship within TransCelerate initiative "Intelligent Automation Opportunities in Pharmacovigilance". His current focus is on Agile ALM class tools and CSA.