by Personalized T cell therapies for cancer patients have long been a complex and time-consuming process, typically taking at least six months to develop. However, researchers at the German Cancer Research Center (DKFZ) and the University Medical Center Mannheim have made a groundbreaking discovery that could significantly accelerate this process. By utilizing a machine learning classifier, the arduous task of identifying tumor-reactive T cell receptors for patients can be streamlined, cutting the development time in half.
Cellular immunotherapies, specifically T-cell receptor transgenic T-cells, are at the forefront of innovative treatment options for various types of cancer. These therapies involve equipping a patient’s immune T cells in the laboratory to target and destroy the patient’s unique tumor cells effectively. The initial step in this process involves isolating tumor-infiltrating T cells (TILs) from a patient’s tumor tissue and identifying T-cell receptors that can recognize and eliminate tumor cells bearing specific mutations.
This search for tumor-reactive T cell receptors has traditionally been a labor-intensive and time-sensitive process, relying on knowledge of the specific tumor mutations recognized by the immune system. With tumors evolving and spreading rapidly, the race against time has been a significant challenge in this process. However, the recent development of a machine learning classifier by a research team led by Michael Platten and Ed Green has proven to revolutionize this aspect of cellular immunotherapy.
The team’s innovative technology leverages single-cell sequencing to characterize TILs from a melanoma patient, followed by individual testing of T cell receptors to identify those capable of targeting and killing tumor cells. By training a machine learning model with this data, the researchers successfully created a classifier that could predict tumor-reactive T cell receptors with an impressive 90% accuracy. Notably, this classifier is versatile and can be applied to various types of tumors, accommodating data from different cell sequencing technologies.
This breakthrough technology, known as PredicTCR, has the potential to reduce the time it takes to identify personalized tumor-reactive T cell receptors from over three months to just a few days, irrespective of tumor type. “PredicTCR enables us to cut the time it takes to identify personalized tumor-reactive T cell receptors from over three months to a matter of days, regardless of tumor type,” explained Ed Green.
Looking ahead, the researchers are dedicated to transitioning this technology into clinical practice in Germany. To support further development and implementation, the biotech start-up Tcelltech has been founded, with PredicTCR serving as a key technology of this new DKFZ spin-off. The impact of this machine learning classifier on accelerating the development of cellular immunotherapies signifies a significant step forward in the fight against cancer.
*Note:
1. Source: Coherent Market Insights, Public sources, Desk research
2. We have leveraged AI tools to mine information and compile it.
