How to solve a problem like Kymriah

In its full-year financials, Novartis announced plans to transform its cell therapy manufacturing process through greater investment and utilisation of digital technologies.

Its end of the year report saw Novartis outlining its five strategic priorities for the future, one of which was entitled ‘Embrace operational excellence’, and within this announced it planned to transform its manufacturing network.

The company’s difficulties in effectively manufacturing its chimeric antigen receptor (CAR)-T therapy, Kymriah (tisagenlecleucel), are public knowledge, after Liz Barrett, head of Novartis Oncology, confirmed to BioPharmaDive​ that the company has encountered difficulties hitting label specifications.

A spokesperson for Novartis told us the company has adapted to the challenges it is facing by changing its investment strategy away from ‘more traditional production technologies’.

“[To focus on] advanced and transformative manufacturing platforms to support Novartis’ product portfolio move from high-volume blockbusters at the end of their life cycle to smaller volume, customised and personalised medicines - e.g. Kymriah,”​ the spokesperson explained.

Data and digital​

The spokesperson outlined that Novartis is looking to achieve this aim by focusing on ‘data and digital technologies.’

Data and digital technologies are increasingly referenced within the industry, especially as the interest in ‘Industry 4.0’ techniques begin to attract greater attention for creating manufacturing efficiencies.

A representative from GSK recently told a conference audience​ that the company had managed to make savings of £80m ($102m) by employing digital technologies associated with the movement, such as big data and artificial intelligence.

When asked how Novartis would be applying digital and data to its manufacturing networks, the spokesperson explained the transition that is underway in its internal processes:

“Currently, most pharmaceutical manufacturing is conducted in a make, test and release process. This means that information about the product isn’t known until near the end of the process.​

“We are currently exploring ways to integrate and analyse data from online sensors that will allow us to monitor conditions of a product batch in real time and predict the outcome of the manufacturing process. Our aim is to reduce quality related write-offs to save cost and reduce manufacturing time,”​ the spokesperson continued.

Artificial intelligence​

For Novartis, investment in new technologies has extended into a cornerstone of any discussion on the potential for digital transformation, artificial intelligence.

The spokesperson acknowledged that the company is still “exploring where best to apply AI in the manufacturing process”​, but noted it has seen promising initial feedback in certain areas.

Particularly, the spokesperson confirmed the potential of the technology to predict equipment failures and the likelihood that a batch of product will expire before it can be sold.

In addition, the spokesperson stated, “[Novartis is] exploring ways for AI to help us find opportunities in our data to make more efficient purchases of materials used during the manufacturing process.”​

Further than this, the company is also looking at how AI can be used for the planning, monitoring, and distribution of medicine to patients.

Data-driven decisions​

In terms of how the company sees digital transforming the company’s manufacturing, the spokesperson concluded that sees the technology enabling ‘data-driven decision-making’. This extends across the business, where it is exploring more ‘experimental projects’, such as using virtual/augmented reality to engage its employees on projects and to use apps to track materials through its supply chain.

All of these changes have arrived alongside more traditional approaches from the pharma giant, which has used its financial resources to expand its in-house CAR-T manufacture capabilities by acquiring its manufacturing partner at the start of the year​ and by investing in building out its own network​.