Skip to main content
Home » Patient care » AstraZeneca: Where data and AI impact lives

AstraZeneca is a global, science-led biopharmaceutical company that focuses on the discovery, development and commercialisation of prescription medicines in Oncology, Rare Diseases and BioPharmaceuticals, including Cardiovascular, Renal and Metabolism as well as Respiratory and Immunology.

“At AstraZeneca, we’re using technology, data and AI to transform our company, accelerate our innovative science and maximise the impact for patients,” explains Cindy Hoots, Chief Digital Officer and CIO. “As a company that thrives on innovation, we are constantly evolving our ways of working and we have embedded AI broadly across the organisation.”

Not only has the company invested in connecting diverse data sources and leveraging AI for new insights, but it is also exploring the application of rapidly evolving technologies.

“We are excited about how generative AI, advanced augmented reality, integrated digital twins and next-generation infrastructure can help us achieve our company’s long-term ambitions by accelerating timelines, increasing the probability of success, improving our employee experience and providing more equitable access to healthcare,” Hoots continues.

Gaining better understanding of disease

Hoots and her team enable scientists to leverage AI to advance understanding of disease biology and ultimately uncover novel drivers of diseases the company aims to treat, prevent and, in the future, cure. “We imagine a world where medicines are discovered and brought to patients by blending human ingenuity with advanced AI and other emerging technologies,” says Hoots.

For example, AstraZeneca scientists use knowledge graphs to make better decisions earlier with less effort and material in labs. Knowledge graphs are networks of contextualised scientific data facts — such as genes, proteins, diseases and compounds — and the relationship between them.

“Our teams use advanced machine learning and AI approaches like GraphML and transformers to generate novel insights into target discovery, biomarker identification, patient stratification and drug response,” says Ben Sidders, Executive Director of Oncology Data Science. “We’ve already seen an AI-guided approach bear fruit with our first disease models based on knowledge graphs focused on understanding drug resistance.”

“We recently moved to the next level, adding a generative AI interface to our knowledge graphs and other solutions so that scientists can query trusted, reliable information using plain language and get instant results at their fingertips, as well as help generate and summarise text,” explains Anna Berg Åsberg, Vice President, R&D IT.

“Once we have a disease target in our sights, our scientists can use AI to determine what molecule to create,” she says. Åsberg’s research colleagues now augment traditional drug design with sophisticated computational methods to predict what molecules to make next and how to make them.

“In the past, this process involved making and testing thousands of small molecules over several years to achieve the right drug properties,” adds Åsberg. Now, AI-enabled processes are impacting both the company’s small molecule and large molecule research.

In antibody discovery, the company can use machine learning-enabled deep screening technology to successfully identify early biologic drug hits in just days compared to traditional discovery methods, which take several months.

Advancing personalised medicine

Whether assessing cough recordings, analysing lung tissue samples or identifying which patients are best suited to participate in a clinical trial, AI tools — validated by human experts — help uncover new precision medicine insights.

“Our data scientists are building machine learning algorithms to combine diverse datasets — such as clinical trial data and real-world data — to identify patterns in disease progression and patient response,” says Jim Weatherall, Vice President, Data Science and AI. “These findings inform the company’s clinical trial designs.”

With the aid of AI-powered tools, medical images serve as a valuable source of data, empowering us to gain deeper insights into how a patient’s distinct genetic composition can impact their response to AstraZeneca treatments.

For instance, AstraZeneca experts are using AI in tumour image analysis to improve the accuracy of assessing tumour volumes from computed tomography (CT) scans. “Today, the process is manual, but this new approach would speed up the ability for radiologists to relay information to clinicians about whether a drug has an effect on tumours,” Åsberg adds.

AI has the potential to improve health outcomes
for people, especially those who live in under-
resourced healthcare systems.

Maximising operation efficiencies and growth

Once an AstraZeneca medicine is authorised for sale by regulatory authorities, AI can help to drive supply chain efficiencies. For instance, at one of AstraZeneca’s largest global sites in Sweden, the company manufactures over 12 billion tablets and capsules each year. “Here, we use AI-powered digital twins that can leverage multiple data sources simultaneously to optimise production schedules,” explains Gurinder Kaur, Vice President of Operations and Enabling Units IT, AstraZeneca.

The technology has delivered a 75% reduction in planning lead time, meaning one person can develop a dispensing plan in only 15 minutes. “This used to take eight hours,” says Kaur. “Our ambition is to leverage many of these tools and apply them throughout our manufacturing and supply network.”

Earlier disease detection and diagnosis

AI has potential to improve health outcomes for people, especially who live in under-resourced healthcare systems. AstraZeneca collaborates with — developers of deep learning algorithms for the interpretation of radiology images — to enhance early-stage lung cancer risk identification in low and middleincome countries.

The partnership capitalises on interpreting routine chest X-rays, which happen in large numbers, for risk of malignant lung nodules which can then be referred for further diagnostic testing, potentially leading to earlier stage diagnosis of lung cancer.

“Ultimately, we hope to support early lung cancer detection, reducing mortality rates and enhancing patient outcomes,” explains Kevin Sirjuesingh, Vice President, Commercial IT, AstraZeneca.

Attracting talent, where data and AI impact lives

Working across geographic boundaries, the biopharmaceutical giant aims to bring together experts from around the world and solve complex problems faster than ever before.

“People are our greatest asset. With 50% female representation in senior technology roles, we are committed to building diverse, inclusive teams that are continuously learning together,” says Åsberg.

“We foster collaborative relationships among IT professionals, data scientists, bioinformaticians, AI engineers and bench scientists so that we can talk about science and tech at pace — all working to the same goal of unlocking what science can do,” explains Weatherall.

“We recognise that humans have unique skills that machines do not, and we focus on investing in our people as much as investing in the technology.” AstraZeneca offers a range of opportunities for fresh graduates and those entering the industry, including graduate student rotations, internships and postdoctoral fellowships.

External collaborations are also key to helping answer big questions in AI. “We start with the challenge we need to solve and identify the best partners, whether academic, tech or industry,” Åsberg adds.

Ethical considerations in place

Transforming the future of healthcare by unlocking the power of what AI and science can do — for people, society and the planet — requires trust.

AstraZeneca was one of the first pharmaceutical companies to develop a list of principles for ethical data and AI use in 2020. They revolve around being explainable and transparent; fair and accountable; human-centric; socially beneficial; private and secure.

“We’ve taken a holistic approach to these new emerging technologies, to ensure we are first addressing the ethical, data privacy, legal and procurement considerations,” says Hoots. “We aim to ensure trust in these solutions and deliver value for our business and for patients.”

Next article