Omnia Health is part of the Informa Markets Division of Informa PLC

This site is operated by a business or businesses owned by Informa PLC and all copyright resides with them. Informa PLC's registered office is 5 Howick Place, London SW1P 1WG. Registered in England and Wales. Number 8860726.

Gen AI may power the next generation of immunotherapies

Article-Gen AI may power the next generation of immunotherapies

Generative AI
A ChatGPT-inspired generative AI system has successfully designed a new immunotherapy drug for cancer. What does this mean for drug discovery?

In the relentless global battle against cancer, generative artificial intelligence (Gen AI) has emerged as a powerful tool in oncology’s arsenal. With the capacity to analyse vast datasets, simulate complex molecular interactions, and tailor treatments to individual patients, Gen AI is reshaping the landscape of immunotherapy — an essential frontier in cancer treatment.  

In a groundbreaking accomplishment, British biotech company Etcembly has tapped into the potential of a ChatGPT-like Gen AI tool to create a cutting-edge immunotherapy drug to treat cancer.   

The new drug ETC-101 falls under the category of bispecific T-cell engagers. It derives its power from the natural T-cell receptors (TCRs), which act as molecular couriers, directing immune 'killer' T-cells towards cancer cells with the intention of annihilating them. However, a glaring issue arises when it comes to natural TCRs — their weak affinity for cancer cells and their unfortunate ability to recognise targets in healthy cells. The solution, according to Ectembly, is extensive molecular engineering to fashion TCRs with greater sensitivity and selectivity to function effectively as drugs. 

RelatedInsurtech chatbots: Synergised digital solution for enhancing patient experience

Central to this discovery is EMLy, an advanced supercomputer armed with the latest machine-learning algorithms. EMLy scours through massive datasets, decoding the intricate 'language' of TCRs and identifying the optimal receptor for a specific target. Following this, a generative large language model (LLM), akin to ChatGPT, steps in to 'rewrite' the genetic code of the TCR, ensuring it achieves maximum potency. 

Michelle Teng - Co-founder and CEO, Etcembly

Michelle Teng, Co-founder and CEO, Ectembly 

The culmination of this journey takes place in the laboratory, where the newly crafted genetic code is put to the test, creating real-life TCR-based drugs, of which ETC-101 stands as a prime example. This therapy has been tailored to target PRAME, a molecule that pervades many cancers known for its very low survival rates. ETC-101's lead drug binds to PRAME with a million-fold greater affinity than natural TCRs, all while bypassing healthy cells. This suggests that the treatment is set to be incredibly potent with minimal side effects. What is even more astonishing is the speed at which this achievement has unfolded, with ETC-101's development taking a mere 11 months, a remarkable contrast to the conventional two-year timeline for TCR discovery and engineering processes. 

“Etcembly was born from our desire to bring together two concepts that are ahead of the scientific mainstream — TCRs and generative AI — to design the next generation of immunotherapies,” says co-founder and CEO, Michelle Teng. “This is an area that’s close to my heart because I suffer from an autoimmune disease myself, and I undergo immunotherapy.”  

RelatedBeyond the hype: Generative AI and healthcare in the Middle East

The hype around Gen AI is vast and well-deserved, but concerns remain over its suitability in healthcare, mostly around its future impact and vulnerability to hidden bias in data sets. ChatGPT provides responses centred on what is most likely to come next based on the input it receives, which can be problematic, according to augmented intelligence firm Squirro’s CEO, Dr. Dorian Selz.  

“ChatGPT is a ‘stochastic’ model, as it generates responses based on probability rather than determinism. But it is limited by the data on which it was trained, which can be problematic,” he says. "This means it can lack context and deliver results that do not make sense, are irrelevant, or are not compliant with enterprise codes. There are legitimate questions about which data is used to produce content and how users' access rights to information are correctly managed. Generative AI can be a powerful tool, but to add genuine and tangible value, it must be integrated into a semantic enterprise search engine and trained on internal data.” 

On that note, Ectembly’s Teng agrees that existing TCR trials are based on a predominantly European population. That’s a limitation and extremely short-sighted, she says, pointing to the wider issue of collating diverse data sets. Gen AI runs on limited data and therefore exhibits bias. This means that the very tool that may be called out for its bias can be used to address bias in healthcare. “We are a very cosmopolitan society. We need a platform like generative AI to scale the discovery. We want to make it tailorable for other ethnic groups. This is the main ethos of our company — it is immunotherapy for everyone; not one size fits all.”  

The next step for the company is comprehensive lab testing before it can enter the market. Teng is planning to initiate clinical trials as early as 2025 while working on new immunotherapy drugs to tackle melanoma, lung cancer, and autoimmune diseases like arthritis. The company is also preparing for the next round of investment fundraising.  

Etcembly's transformative AI technology is shattering the barriers that have long impeded the discovery and engineering of TCR candidates. It’s an example of how Gen AI could speed up drug discovery and delivery, ushering in a new era of medical progress. 


Access the must-read eBook on AI here to discover perspectives on its transformative role in the healthcare industry.

Back to Technology

Hide comments


  • Allowed HTML tags: <em> <strong> <blockquote> <br> <p>

Plain text

  • No HTML tags allowed.
  • Web page addresses and e-mail addresses turn into links automatically.
  • Lines and paragraphs break automatically.